Search Images Maps Play YouTube News Gmail Drive More »
Sign in
Screen reader users: click this link for accessible mode. Accessible mode has the same essential features but works better with your reader.

Patents

  1. Advanced Patent Search
Publication numberUS20070044008 A1
Publication typeApplication
Application numberUS 11/527,676
Publication dateFeb 22, 2007
Filing dateSep 27, 2006
Priority dateApr 15, 2002
Also published asUS7127667, US7581160, US20030196163
Publication number11527676, 527676, US 2007/0044008 A1, US 2007/044008 A1, US 20070044008 A1, US 20070044008A1, US 2007044008 A1, US 2007044008A1, US-A1-20070044008, US-A1-2007044008, US2007/0044008A1, US2007/044008A1, US20070044008 A1, US20070044008A1, US2007044008 A1, US2007044008A1
InventorsHong-Ching Chen, Wen-Zen Shen, Der-Tsuey Shen Wang
Original AssigneeMediatek, Inc.
Export CitationBiBTeX, EndNote, RefMan
External Links: USPTO, USPTO Assignment, Espacenet
ACS circuit and Viterbi decoder with the circuit
US 20070044008 A1
Abstract
An ACS circuit and a Viterbi decoder with the circuit. The Add-Compare-Select (ACS) circuit comprises: two registers for storing two previous candidate state metrics; a first adder for adding the value stored in the first register and a first branch metric to generate a first addition result; a second adder for adding the value stored in the second register and the first branch metric to generate a second addition result; a comparator for comparing the values stored in the first register and the second register to generate a decision bit; and a multiplexer for selecting either the first addition result or the second addition result as a new output candidate state metric according to the decision bit. Due to the parallel processing of the adders and the comparator, the processing speed of a Viterbi decoder with the ACS will be increased.
Images(10)
Previous page
Next page
Claims(8)
1. A Viterbi decoder for decoding a string of data, a trellis structure in the Viterbi decoder containing N trellis states with two-branches and K trellis states with only one branch, wherein N and K are positive integers, the Viterbi decoder comprising:
a branch metric generator for calculating and outputting branch metrics for each branch;
an add-compare-select unit for receiving the branch metrics and previous candidate state metrics respective to the trellis state in the trellis structure and outputting N comparison signals;
wherein the add-compare-select unit comprises:
(2N+K) registers for storing the previous candidate state metrics respective to the trellis state;
N add-compare-select processors for receiving values stored in the 2N first registers and corresponding branch metrics, and outputting N candidate state metrics and the N comparison signals; and
wherein each add-compare-select processor comprises:
a first adder for adding a first value stored in one of the registers and a first branch metric to generate a first addition result;
a second adder for adding a second value stored in another of the registers and the first branch metric to generate a second addition result;
a comparator for comparing the first value and the second value, and outputting a comparison signal; and
a selector for selecting either the first addition result or the second addition result as the candidate state metric according to the comparison signal;
wherein each add-compare-select processor is configured with the registers being located before the comparator.
2. The Viterbi decoder of claim 1, wherein the Viterbi decoder is used for decoding the string of data using partial response maximum likelihood with run-length code constraint in an optical storage system.
3. A Viterbi decoder for decoding a string of data, a trellis structure in the Viterbi decoder containing N trellis states with two-branches and K trellis states with only one branch, wherein N and K are positive integers, the Viterbi decoder comprising:
a branch metric generator for calculating and outputting branch metrics for each branch;
an add-compare-select unit for receiving the branch metrics and previous candidate state metrics respective to the trellis state in the trellis structure and outputting N comparison signals;
wherein the add-compare-select unit comprises:
(2N+K) registers for storing the previous candidate state metrics respective to the trellis state;
N add-compare-select processors for receiving values stored in the 2N first registers and corresponding branch metrics, and outputting N first candidate state metrics, N second candidate state metrics and the N comparison signals;
wherein each add-compare-select processor comprises:
a first adder for adding a first value stored in one of the registers and a first branch metric to generate a first addition result;
a second adder for adding a second value stored in another of the registers and the first branch metric to generate a second addition result;
a third adder for adding the first value and a second branch metric to generate a third addition result;
a forth adder for adding the second value and the second branch metric to generate a fourth addition result;
a comparator for comparing the first value and the second value, and outputting a comparison signal;
a first selector for selecting either the first addition result or the second addition result as the first candidate state metric according to the comparison signal; and
a second selector for selecting either the third addition result or the fourth addition result as the second candidate state metric according to the comparison signal;
wherein each add-compare-select processor is configured with the registers being located before the comparator.
4. The Viterbi decoder of claim 3, wherein the Viterbi decoder is used for decoding the string of data using partial response maximum likelihood with run-length code constraint in an optical storage system.
5. A decoding method of a Viterbi decoder for decoding a string of data, a trellis structure in the Viterbi decoder containing N trellis states with two-branches and K trellis states with only one branch, wherein N and K are positive integers, the Viterbi decoder having a branch metric generator and an add-compare-select unit, the add-compare-select unit having (2N+K) registers and N add-compare-select processors, the method comprising the steps of:
calculating and outputting branch metrics for each branch by the branch metric generator;
adding 2N previous candidate state metrics respective to a trellis state in the trellis structure and a first branch metric from the trellis state, respectively, to generate N first addition results and N second addition results by the add-compare-select processor;
comparing the 2N previous candidate state metrics, and outputting N comparison signals by the add-compare-select processor;
selecting N candidate state metrics from the N first addition results and the N second addition results according to the N comparison signals by the add-compare-select processor;
wherein the adding step and the comparing step are executed concurrently.
6. The decoding method of claim 5, wherein the decoding method is used for decoding the string of data using partial response maximum likelihood with run-length code constraint in an optical storage system.
7. A decoding method of a Viterbi decoder for decoding a string of data, a trellis structure in the Viterbi decoder containing N trellis states with two-branches and K trellis states with only one branch, wherein N and K are positive integers, the Viterbi decoder having a branch metric generator and an add-compare-select unit, the add-compare-select unit having (2N+K) registers and N add-compare-select processors, the method comprising the steps of:
calculating and outputting branch metrics for each branch by the branch metric generator;
adding 2N previous candidate state metrics respective to a trellis state in the trellis structure and a first branch metric from the trellis state, respectively, to generate N first addition results and N second addition results, and adding the 2N previous candidate state metrics and a second branch metric from the trellis state, respectively, to generate N third addition results and N fourth addition results by the add-compare-select processor;
comparing 2N previous candidate state metrics, and then outputting N comparison signals by the add-compare-select processor;
selecting N first candidate state metrics from the N first addition results and the N second addition results, and selecting N second candidate state metrics from the N third addition results and the N fourth addition results according to the N comparison signals by the add-compare-select processor; and
wherein the adding step and the comparing step are executed concurrently.
8. The decoding method of claim 7, wherein the decoding method is used for decoding the string of data using partial response maximum likelihood with run-length code constraint in an optical storage system.
Description
  • [0001]
    This application is a Continuation of co-pending application Ser. No. 10/412,339, filed on Apr. 14, 2003, the entire contents of which are hereby incorporated by reference and for which priority is claimed under 35 U.S.C. 120.
  • BACKGROUND OF THE INVENTION
  • [0002]
    1. Field of the Invention
  • [0003]
    The invention relates to an ACS (Add-Compare-Select) circuit, and in particular to an ACS circuit, in which an adder and a comparator are processing in parallel to increase the processing speed thereof, thereby increasing the processing speed of a Viterbi decoder with the ACS circuit.
  • [0004]
    2. Description of the Related Art
  • [0005]
    A PRML (Partial Response Maximum Likelihood) system is used to retrieve EFM (Eight-to-Fourteen Modulation) signals from CD (Compact Disc)/DVD (Digital Versatile Disk). In the PRML system, a Viterbi decoder is usually used to realize the maximum likelihood detection.
  • [0006]
    FIG. 1(A) is a block diagram showing a conventional Viterbi decoder. Referring to this drawing, a Viterbi decoder typically includes a branch metric generator 11, an ACS (Add-Compare-Select) unit 12, and a survivor path memory and decoding unit 13. The branch metric generator 11 calculates the metrics corresponding to each branch according to the received data, and outputs them to the ACS unit 12. The branch metric generator 11 calculates the distance, such as mean-square-distance, between the received data and the estimated data of the branch. For each state, the ACS unit 12 adds the branch metrics of input branches with the related state metric to generate the candidate state metrics, compares and selects the minimum of the candidate state metrics as the new state metric. The decision bit of the ACS unit 12, indicating the decision of the state-metric selection, is sent to the survivor path memory and decoding unit 13. According to the decision bit, the survivor paths of the survivor path memory and decoding unit 13 are updating to keep consistent with the selected result of related state metric. The survivor memory and decoding unit 13 restores the decoded results corresponding to each state, and a decoding unit executes the majority vote for the ending bits of all survivor paths to decide the output decoded result.
  • [0007]
    FIG. 1(B) is an architecture diagram showing an ACS processor in the ACS unit of FIG. 1(A), wherein SMi(n−1) and SMj(n−1) are state metrics of state Si and state Sj at time n−1, respectively, and BMj,k(n−1) and BMj,k(n−1) are branch metrics from state Si to state Sk at time n−1 and from state Sj to state Sk at time n−1, respectively. The SMk(n) is a state metric at time n and the BMk,m(n) is the branch metric from state Sk to state Sm at time n. In this drawing, it illustrates that the ACS processor is used to select the survivor state metric from either state Si or state Sj. As shown in the drawing, an ACS processor 14 in an ACS unit 12 includes two adders 15, 15′, a comparator 16, and a selector 17. The ACS processor 14 employs the adder 15 to add SMi(n−1) to BMi,k(n−1) to generate a first candidate state metric, and the adder 15′ to add SMj(n−1) to BMj,k(n−1) to generate a second candidate state metric. Then, the ACS processor 14 employs the comparator 16 and selector 17 to compare the first candidate state metric with the second candidate state metric, and select the minimum to be outputted to a state register 18 as the SMk(n). The comparison result, serving as a decision bit, is outputted to the survivor memory and decoding unit 13. The adder 19 belongs to another ACS processor 14′.
  • [0008]
    FIG. 2(A) shows a 4-state trellis diagram, and FIG. 2(B) is a block diagram showing the ACS unit 12 in the Viterbi decoder corresponding to the trellis diagram of FIG. 2(A). Since FIG. 2(A) is a 4-state trellis diagram, the ACS unit 12 includes four ACS processors 121 to 124 and four state registers 125 to 128 for storing the state metrics S0 to S3, as shown in FIG. 2(B). In addition to transferring the generated metrics back to the state registers 125 to 128, the ACS processors 121 to 124 of the ACS unit 12 further output decision bits to the survivor memory and decoding unit 13. The ACS processor 121 receives the state metrics from the state registers 125 and 127, adds together the state metrics and the related branch metrics, selects a smaller addition result as a survivor metric, stores back to the state register 125, and outputs the comparison value (logic 1 or logic 0) serving as a decision bit simultaneously. The ACS processor 122 receives the state metrics from the state register 125 and 127, adds together the state metrics and the related branch metrics, selects a smaller addition result as a survivor metric, stores back to the state register 126, and outputs a comparison value (logic 1 or logic 0) serving as a decision bit simultaneously. The processing methods for the ACS processors 123 and 124 are also the same.
  • [0009]
    FIG. 3 shows a circuit of the ACS processor as shown in FIG. 2(B). Referring to FIG. 3, the ACS processor 121 includes two adders 1211, 1211′, a comparator 1212 and a multiplexer 1213. The adders 1211, 1211′ are used to add together the branch metric and the state metric, and the comparator 1212 is used to compare two output values from the adders 1211, 1211′. The multiplexer 1213 selects a value from the output values of the two adders 1211, 1211′ as a new state metric according to the comparison result of the comparator 1212. Meanwhile, the comparison result of the comparator 1212 also serves as a decision bit for output. From FIG. 1(B) and FIG. 3, it is shown that the execution order of the ACS processor is Add->Compare->Select. Because the execution of comparison needs to wait the result of addition, the addition and comparison cannot be executed in parallel, which is a timing bottleneck of a conventional ACS unit.
  • [0010]
    In the application of decoding an EFM signal, the EFM signal has the property of run length limited (RLL) that the minimum run length of an EFM signal is 3 T, where T is a recording unit length. When decoding with respect to the EFM signal, a simplified trellis diagram can be obtained according to the RLL of the EFM signal, as shown in FIG. 4(A). That is, the trellis diagram has six states including state S0(000), state S1(001), state S2(011), state S3(100), state S4(110), and state S5(111), in which the states of (010) and (101) are invalid and not listed due to the minimum run length of 3 T. In addition, state S1(001), state S2(011), state S3(100) and state S4(110) only have a branch. FIG. 4(B) is a schematic illustration showing the branch values corresponding to the trellis diagram of FIG. 4(A). The Viterbi decoder generates all possible input sequences from the trellis diagram, and selects the most possible result as the decoded result. According to the trellis diagram of FIG. 4(A), only state S0 and state S5 receive two input values, so the state S0 and state S5 have to judge and select one of the input sequences.
  • [0011]
    FIG. 5 is an architecture diagram showing a Viterbi decoder applied to the trellis diagram of FIG. 4(A). Referring to the drawing, the Viterbi decoder includes an ACS unit 52 and a survivor memory and decoding unit 13. Because the trellis diagram has already been simplified due to the property of RLL, the ACS unit 52 only includes two ACS processors 521 and 522, two adders 523 and 524, and six registers 525 to 530. The ACS processor 521 receives the state metrics of state S0 and state S3, adds together the state metrics and the related branch metrics, selects a smaller result through the compare-select circuit 5211, and stores the selected result back to the register 525 at state S0. Next, the ACS processor 521 outputs the comparison value, serving as a decision bit, from the compare-select circuit 5211 to the multiplexers in the path 0 of the survivor memory and decoding unit 13. The ACS processor 522 receives the state metrics of state S2 and state S5, adds together the state metrics and the branch metrics, selects a smaller result through the compare-select circuit 5221, and stores the selected result back to the register 530 at state S5. Next, the ACS processor 522 outputs the comparison value, serving as a decision bit, from the compare-select circuit 5221 to the multiplexers in the path 5 of the survivor memory and decoding unit 13. Since the trellis diagram has been simplified, the multiplexers are only arranged in path 0 and path 5 of the survivor memory and decoding unit 13. Other paths 1 to 4 are only used to transfer data directly to another path memory. The survivor memory and decoding unit 13 employs a decision circuit 131, which may be a majority vote circuit, to vote the majority of bit value of the ending bits of six survivor paths as decoded data for output.
  • [0012]
    U.S. Pat. No. 6,148,431 titled “Add compare select circuit and method implementing a Viterbi algorithm” disclosed an ACS circuit with parallel process in the adder and comparator. In this architecture, the trellis diagram must satisfy the condition that the branch metrics for the ACS unit have to be equal. That is, the ACS unit in the U.S. Pat. No. 6,148,431 is not an ACS unit for general purpose.
  • [0013]
    In general, the conventional ACS unit is the bottleneck of processing speed of the Viterbi decoder, and the ACS processor of the ACS unit cannot increase the processing speed by way of direct pipelining or parallel processing.
  • SUMMARY OF THE INVENTION
  • [0014]
    In view of the above-mentioned problems, an object of the invention is to provide a Viterbi decoder capable of changing the processing order of the ACS processor to make the add and compare units of the ACS processor process in parallel, thereby increasing the processing speed of the Viterbi decoder.
  • [0015]
    To achieve the above-mentioned object, the invention provides a Viterbi decoder consisting of a branch metric generator, a plurality of ACS units, and a survivor memory and decoding unit. The ACS unit comprises: two registers for storing two previous candidate state metrics; and a ACS processor for receiving the previous candidate state metrics stored in the state registers and a first branch metric, generating a first output candidate state metrics.
  • [0016]
    The ACS processor comprises a first adder, a second adder, a comparator and a first multiplexer. The first adder adds the value stored in the first register and a first branch metric to generate a first addition result. The second adder adds the value stored in the second register and the first branch metric to generate a second addition result. The comparator compares the values stored in the first register and the second register, and outputs a comparison signal as the decision bit. The first multiplexer selects either the first addition result or the second addition result as the first output candidate state metric according to the comparison signal.
  • [0017]
    The ACS processor can further comprise a third adder, a fourth adder, and a second multiplexer. The third adder adds the value stored in the first register and a second branch metric to generate a third addition result. The fourth adder adds the value stored in the second register and the second branch metric to generate a fourth addition result. The second multiplexer selects either the third addition result or the fourth addition result as a second output candidate state metric according to the comparison signal of the comparator.
  • [0018]
    Due to the parallel processing of the adders and the comparator of the ACS processor, the processing speed of the Viterbi decoder is increased.
  • BRIEF DESCRIPTION OF THE DRAWINGS
  • [0019]
    FIG. 1(A) is a block diagram showing a conventional Viterbi decoder.
  • [0020]
    FIG. 1(B) is an architecture diagram showing the ACS processor in the ACS unit of FIG. 1(A).
  • [0021]
    FIG. 2(A) shows an example of the 4-state trellis diagram.
  • [0022]
    FIG. 2(B) is a block diagram showing the ACS unit of the Viterbi decoder corresponding to the trellis diagram of FIG. 2(A).
  • [0023]
    FIG. 3 shows the circuit of the ACS processor of FIG. 2(B).
  • [0024]
    FIG. 4(A) shows a simplified trellis diagram obtained according to the RLL of the EFM signal.
  • [0025]
    FIG. 4(B) is a schematic illustration showing the branch values corresponding to the trellis diagram of FIG. 4(A).
  • [0026]
    FIG. 5 is an architecture diagram showing a typical Viterbi decoder applied to the trellis diagram of FIG. 4(A).
  • [0027]
    FIG. 6 is an architecture diagram showing the ACS processor in which the operation order is rearranged.
  • [0028]
    FIG. 7 shows the circuit of the ACS processor of the invention, wherein FIG. 7(A) shows an adder arranged in back of a multiplexer, while FIG. 7(B) shows two adders arranged in front of a multiplexer.
  • [0029]
    FIG. 8 shows the circuit of the PACS unit related to a state with two output branches.
  • [0030]
    FIG. 9 is a block diagram showing an ACS unit of the Viterbi decoder according to a first embodiment of the invention.
  • [0031]
    FIG. 10(A) shows a data flow diagram of the typical ACS unit according to the trellis diagram of FIG. 4(A).
  • [0032]
    FIG. 10(B) shows a data flow diagram of the ACS unit of the present invention according to the trellis diagram of FIG. 4(A).
  • DETAILED DESCRIPTION OF THE INVENTION
  • [0033]
    The Viterbi decoder of the invention will be described with reference to the accompanying drawings.
  • [0034]
    First, please refer back to FIG. 1(B). In the architecture of the ACS processor of a conventional Viterbi decoder, an adding process is firstly executed, and then a compare-select unit is utilized to select the corresponding data to be outputted to the register. That is, the operation order is add->compare->select. Consequently, parallel processing of the adder and comparator can not be directly executed in this architecture. In view of this, the invention rearranges the operation order of the ACS processor to make the adder and comparator process in parallel to increase the processing speed of the ACS processor.
  • [0035]
    FIG. 6 is an architecture diagram showing an ACS processor of the present invention in which the operation order is rearranged. Referring to FIG. 6, the ACS processor 60 includes a comparator 63, a selector 65 and an adder 64. Compared with FIG. 1(B), the present invention moves the state metric register forward from the location behind the selector into the location before the comparator, which has to add an extra state metric register. That is, each ACS unit in the present invention needs more state metric register. The comparator 63 receives the signals from the two state metric registers 62, 62′ and outputs a decision bit. Then the selector 65 selects one of the signals from the two state metric registers 62, 62′ according to the decision bit and outputs a new state metric. Finally, the adder 64 adds the state metric outputted from the selector 65 with a branch metric and outputs the result to another metric registers (not shown). Therefore, the operation order of the ACS processor 60 is changed into compare->select->add, which is different from the conventional operation order of add->compare->select.
  • [0036]
    FIG. 7 shows the circuit of the ACS processor of the invention, wherein FIG. 7(A) shows an adder arranged in back of a multiplexer, while FIG. 7(B) shows two adders arranged in front of a multiplexer. As shown in FIG. 7(A), the ACS processor 70 utilizes the comparator 71 to compare the values of two candidate state metrics (a, b) and to generate a decision bit, and utilizes the multiplexer 72 to select a smaller candidate state metric as a survivor state metric according to the decision bit. The candidate state metrics (a, b) are stored in different state registers, respectively. Then, the adder 73 is utilized to add together the selected survivor state metric and the branch metric (c) to generate a new candidate state metric, which is sent to anther ACS processor. Therefore, the output value 0 of the ACS processor 70 is:
    O=min(a,b)+c  (1).
  • [0037]
    Substituting the branch metric (c) of Equation (1) into the min function, Equation (1) is converted into Equation (2):
    O=min(a+c,b+c)  (2)
  • [0038]
    FIG. 7(B) shows the circuit of the ACS processor generated according to Equation (2). As shown in the drawing, because the adders 73 are moved to positions in front of the multiplexer 72 in the ACS processor 75, an extra adder 73′ is added, and the adders 73, 73′ and comparator 71 can perform parallel processing. Consequently, the ACS processor 75 has the processing speed defined as (delay time of the adder+delay time of the multiplexer), or (delay time of the comparator+delay time of the multiplexer), which is quicker than the processing speed defined as (delay time of the adder+delay time of the comparator+delay time of the multiplexer) of prior art. In general, the delay time of the adder substantially equals to the delay time of the comparator. Therefore, when the bit length of the state metric is so long that the delay time of the multiplexer can be neglected, the processing speed of the ACS processor 75 can be substantially doubled.
  • [0039]
    FIG. 8 shows the circuit of the PACS unit related to a state with two output branches. Compared to FIG. 7, it outputs two candidate state metrics of two output branches. The PACS unit 80 shown in FIG. 8 includes a comparator 81, two multiplexers 82 and four adders 83. That is, the PACS unit 80 includes an extra multiplexer 82 and two extra adders 83 with respect to the PACS unit 75.
  • [0040]
    FIG. 9 is a block diagram showing an ACS unit of the Viterbi decoder according to a first embodiment of the invention. The Viterbi decoder as shown in FIG. 9 will be described taking the 4-state trellis diagram of FIG. 2(A) as an example. The ACS unit 90 includes four PACS (parallel Add-Compare-Select) unit 80 for parallel processing, and eight state registers 92 for storing state metrics. That is, each state branch needs two state registers 92 and a PACS unit 80. The two candidate state metrics outputted from each PACS unit 80 of the ACS unit 90 are outputted to two state registers 92.
  • [0041]
    FIG. 10 shows the data flow diagram of the ACS unit generated according to the trellis diagram of FIG. 4(A), wherein FIG. 10(A) shows the data flow diagram of the conventional ACS unit, while FIG. 10(B) shows the data flow diagram of the ACS unit of the present invention. The bold lines in FIG. 10 denote the main data flow. As shown in the FIG. 10(A), the data from the register D must pass through the adder, the comparator, and then the selector. On the other hand, as shown in FIG. 10(B), the data from the register D just passes through the adder and then the selector. In this embodiment, the hardware cost of the invention is increased to 20%˜30% of the original cost. However, the speed is substantially doubled. Therefore, the invention's ACS unit can effectively shorten the data processing time of the ACS processor, thereby increasing the data processing speed of the ACS unit.
  • [0042]
    While certain exemplary embodiments have been described and shown in the accompanying drawings, it is to be understood that such embodiments are merely illustrative of and not restrictive on the broad invention, and that this invention not be limited to the specific construction and arrangement shown and described, since various other modifications may occur to those ordinarily skilled in the art.
Patent Citations
Cited PatentFiling datePublication dateApplicantTitle
US4612659 *Jul 11, 1984Sep 16, 1986At&T Bell LaboratoriesCMOS dynamic circulating-one shift register
US4630295 *Jul 24, 1984Dec 16, 1986Sharp Kabushiki KaishaLow power consumption CMOS shift register
US5280489 *Apr 15, 1992Jan 18, 1994International Business Machines CorporationTime-varying Viterbi detector for control of error event length
US5430744 *Sep 30, 1993Jul 4, 1995International Business Machines CorporationMethod and means for detecting partial response waveforms using a modified dynamic programming heuristic
US5588011 *Feb 5, 1996Dec 24, 1996C. M. RiggleTuned viterbi detector and equalizer system
US5719843 *Jun 21, 1996Feb 17, 1998Matsushita Electric Industrial Co.LtdMethod of maximum likelihood decoding and digital information playback apparatus
US5729517 *Oct 28, 1996Mar 17, 1998Sharp Kabushiki KaishaData detecting circuit
US5799046 *Jan 23, 1996Aug 25, 1998Pioneer Electric CorporationDigital data signal reproducing apparatus
US5815515 *Mar 21, 1997Sep 29, 1998Lsi Logic CorporationEdge metric calculation method and apparatus using permutations
US5835510 *Jun 10, 1996Nov 10, 1998Pioneer Electronic CorporationHigh-speed viterbi decoder
US5928378 *Aug 22, 1997Jul 27, 1999Daewoo Electronics Co., Ltd.Add-compare-select processor in Viterbi decoder
US5940416 *Jun 24, 1997Aug 17, 1999Hitachi Ltd.Digital signal decoding apparatus and a decoding method used therein
US5987638 *Apr 22, 1997Nov 16, 1999Lsi Logic CorporationApparatus and method for computing the result of a viterbi equation in a single cycle
US6111835 *Jul 24, 1998Aug 29, 2000Nec CorporationPRML decoder for processing different channel codes with reduced hardware
US6148431 *Mar 26, 1998Nov 14, 2000Lucent Technologies Inc.Add compare select circuit and method implementing a viterbi algorithm
US6298464 *Dec 2, 1998Oct 2, 2001Motorola Inc.Method and apparatus for maximum likelihood sequence detection
US6477661 *Oct 12, 2001Nov 5, 2002Matsushita Electric Industrial Co., Ltd.Processing unit and processing method
US6532337 *Aug 19, 1998Mar 11, 2003Sony CorporationDigital-signal playback apparatus
US6667841 *Feb 21, 2001Dec 23, 2003Fujitsu LimitedApparatus for recording and regenerating data
US6738941 *Dec 23, 1999May 18, 2004Nec CorporationData error correction system
US6785212 *Jul 31, 2000Aug 31, 2004Samsung Electronics Co., Ltd.Optical disk reproducing apparatus having a viterbi detector
US6792571 *Nov 2, 2001Sep 14, 2004Acer Laboratories Inc.Viterbi detector for partial response maximum likelihood signal processing
US6798726 *Aug 1, 2001Sep 28, 2004Koninklijke Philips Electronics N.V.Data decoding
US6799296 *Nov 1, 2001Sep 28, 2004Samsung Electronics Co., Ltd.Viterbi detector for optical disk system
US7127667 *Apr 14, 2003Oct 24, 2006Mediatek Inc.ACS circuit and viterbi decoder with the circuit
US20020094057 *Dec 13, 2001Jul 18, 2002Yoshihiro ShibuyaShift register
US20020122480 *Mar 12, 2001Sep 5, 2002Arthur AbnousArchitecture for very high-speed decision feedback sequence estimation
US20020174402 *May 17, 2002Nov 21, 2002Matsushita Electric Industrial Co., Ltd.Viterbi detector
US20030196163 *Apr 14, 2003Oct 16, 2003Hong-Ching ChenACS circuit and viterbi decoder with the circuit
US20040044946 *Aug 30, 2002Mar 4, 2004Bickerstaff Mark AndrewHigher radix Log MAP processor
US20080072129 *Oct 30, 2007Mar 20, 2008Matsushita Electric Industrial Co., Ltd.Communications digital signal processor and digital signal processing method
Referenced by
Citing PatentFiling datePublication dateApplicantTitle
US7728744Sep 13, 2007Jun 1, 2010Analog Devices, Inc.Variable length decoder system and method
US8024551Oct 26, 2005Sep 20, 2011Analog Devices, Inc.Pipelined digital signal processor
US8074157 *Jan 22, 2008Dec 6, 2011Agere Systems Inc.Methods and apparatus for reduced complexity soft-output viterbi detection
US8285972Oct 26, 2005Oct 9, 2012Analog Devices, Inc.Lookup table addressing system and method
US8301990Sep 27, 2007Oct 30, 2012Analog Devices, Inc.Programmable compute unit with internal register and bit FIFO for executing Viterbi code
US8458445Aug 10, 2011Jun 4, 2013Analog Devices Inc.Compute units using local luts to reduce pipeline stalls
US8578255Dec 19, 2008Nov 5, 2013Altera CorporationPriming of metrics used by convolutional decoders
US8914716 *Dec 22, 2008Dec 16, 2014Altera CorporationResource sharing in decoder architectures
US20070094483 *Oct 26, 2005Apr 26, 2007James WilsonPipelined digital signal processor
US20070263753 *Dec 6, 2006Nov 15, 2007Samsung Electronics Co., Ltd.Viterbi decoding method and viterbi decoder
US20080010439 *Sep 13, 2007Jan 10, 2008Yosef SteinVariable length decoder system and method
US20090089649 *Sep 27, 2007Apr 2, 2009James WilsonProgrammable compute unit with internal register and bit FIFO for executing Viterbi code
US20090114111 *Oct 27, 2008May 7, 2009Andreas HeineExplosive charge
US20090187813 *Jan 22, 2008Jul 23, 2009Haratsch Erich FMethods and Apparatus for Reduced Complexity Soft-Output Viterbi Detection
US20090228768 *Dec 22, 2008Sep 10, 2009Altera CorporationResource sharing in decoder architectures
Classifications
U.S. Classification714/795
International ClassificationH03M13/03, H03M13/41
Cooperative ClassificationH03M13/4107, H03M13/6343
European ClassificationH03M13/63P, H03M13/41A
Legal Events
DateCodeEventDescription
Feb 25, 2013FPAYFee payment
Year of fee payment: 4
Feb 27, 2017FPAYFee payment
Year of fee payment: 8